bioRxiv Science⌕ Search

Biology subjects

Liu, H.

Publications and source records attributed to Liu, H..

At least 199 records · Page 11Linked to original sources

Structure based discovery of antipsychotic-like TAAR1 agonists

Schizophrenia is a severe mental illness whose current treatments primarily target dopamine and serotonin receptors. These drugs often cause side effects and vary in effectiveness across patients. The trace amine-associated receptor 1 (TAAR1), which modulates monoamine signaling, has emerged as a promising alternative target. To discover new TAAR1 ligands, we computationally docked 65 million molecules against the active state of TAAR1 and experimentally tested 55 of those highly ranked. Fourteen molecules active against TAAR1 with potencies ranging from mid-nanomolar to micromolar, all as agonists. This high functional selectivity may reflect the compact conformation adopted by the activated TAAR1 orthosteric site. While this was favorable for agonist prioritization, simulations suggest that it can be over-optimized for initial hit rates at the expense of subsequent affinity maturation. Here, hit optimization yielded nanomolar agonists whose docking-predicted poses were confirmed by cryo-EM. Three agonists had high brain exposure and potencies apparently better than the investigational drug ulotaront and sufficient for behavioral studies. All three potently normalized amphetamine-induced pre-pulse inhibition in mice, a model for schizophrenia, without catalepsy, a common side effect of traditional antipsychotics.

biochemistry↗

Single-Cell RNA and ATAC Sequencing Reveal Cellular Heterogeneity and Chromatin Accessibility Dynamics in Young Chinese Breast Cancer Patients across Pre- and Post-Neoadjuvant Therapy

Young breast cancer (YBC) patients (age [≤] 40) exhibit aggressive features and poor prognosis, while facing complex needs including breast-conserving surgery and fertility preservation. Neoadjuvant therapy (NAT) is important for tumor downstaging and breast-conserving surgery in YBC treatment, but the NAT response mechanisms in YBC patients remain unelucidated. Here, we analyzed pre- and post-NAT samples from Chinese YBC patients using scRNA-seq and scATAC-seq. We found that NAT reshaped cellular heterogeneity in the tumor microenvironment (TME), reducing epithelial and T cells while increasing endothelial cells and fibroblasts. Residual cancer cells showed enriched epithelial-mesenchymal transition (EMT) and stromal remodeling programs. NAT responders had fewer luminal-like cells and retained basal-like cells in an early hybrid EMT state, whereas non-responders maintained both populations with late hybrid EMT. We identified therapy-resistant genes and motifs (e.g., VTCN1, PROM1, MZB1, Fox family) and linked CXCL13 upregulation to poor NAT response and tumor-specific T-cell expansion. Cell-cell communication analysis revealed NAT reprograms the TME by suppressing VEGF and TNF signaling in basal cells, while residual luminal cells transmit EGFR and CD47-SIRPA signals.

cancer biology↗

Bioinformatic analysis and experimental validation of ursolic acid's effects on Staphylococcus aureus-induced osteomyelitis

Osteomyelitis (OM), particularly methicillin-resistant Staphylococcus aureus (MRSA)-induced OM, remains a serious clinical challenge due to complex pathogenesis and rising antibiotic resistance. This study aimed to elucidate the therapeutic mechanism of ursolic acid (UA) against MRSA- induced OM using integrated network pharmacology, bioinformatics, and experimental validation. Through target prediction and multi-database mining, we identified 57 common targets of UA and OM. ML algorithms, including Least Absolute Shrinkage and Selection Operator (LASSO) and random forest, pinpointed three core genes: Neutrophil Elastase (ELANE), Lactoferrin (LTF), and S100A12. Enrichment analyses revealed significant involvement in neutrophil extracellular trap (NET) formation, inflammatory response, and immune regulation. Molecular docking and dynamics simulations confirmed stable binding between UA and ELANE. In vitro, UA exhibited antibacterial activity against clinical MRSA isolates (MIC = 16-64 g/mL) and showed synergistic effects with Penicillin (FICI = 0.125-0.5). These results demonstrate that UA combats OM through immunomodulation, anti-inflammatory actions, and direct antibacterial activity, providing a foundation for its use as an adjuvant therapy against MRSA-related OM.

microbiology↗

Fragile nucleosomes are essential for RNA Polymerase II to transcribe in eukaryotes

Nucleosomes are barriers to RNA Polymerase II (Pol II) transcribing along gene bodies in eukaryotes. We found that a fragile "tailless nucleosome" could be generated to resolve the Pol II pausing caused by the +1 nucleosome in metazoans. We hypothesized that "pan-acetylated nucleosomes" along gene bodies are fragile and are generated to resolve the resistance of nucleosomes to the elongation of Pol II. To confirm this hypothesis, we systematically analyzed pan-acetylation and pan-phosphorylation of the entire human genome in HEK293T cells. Herein, we show that both pan-acetylation and pan-phosphorylation of histones in nucleosomes along gene bodies are coupled with the activities of Pol II. Our results show that there are four major types of fragile nucleosomes required for Pol II to transcribe along gene bodies: nucleosomes without tails, nucleosomes with dominant acetylation, nucleosomes with dominant phosphorylation, and nucleosomes with both acetylation and phosphorylation.

genetics↗

Psoralen Enhances Direct Cardiac Reprogramming via PPARα Activation and Promotion of Mitochondrial Fission

BACKGROUNDCell-based therapy is a promising strategy for heart repair and regeneration. However, its clinical application has been hampered by the low efficiency of cell direct reprogramming and the immature state of the inducing cells. Developing novel strategies to enhance direct reprogramming efficiency and yield mature functional cardiomyocytes remains a critical challenge. METHODSWe evaluated the effect of Psr on cardiac reprogramming using RepSox and Forskolin (RF) as the baseline cocktail. Reprogramming efficiency, cardiomyocyte marker expression, and calcium handling were assessed by RT-qPCR, immunofluorescence. Ultrastructural features were examined by transmission electron microscopy. Metabolic profiles were analyzed using Seahorse assays. Transcriptomic changes were explored by RNA sequencing, followed by pathway and protein-protein interaction analyses. A myocardial infarction (MI) mouse model was used for in vivo validation. RESULTSSupplementation with Psr (10 M) markedly enhanced the induction of induced cardiomyocytes (iCMs), leading to earlier appearance of beating clusters (day 1 vs. day 6-8), improved sarcomere organization, robust calcium transients, and higher energy metabolism. Transcriptomic profiling identified activation of the PPAR signaling pathway, with PPAR, RXRG, and UCP1 as central regulators. Mechanistically, Psr promoted mitochondrial fission, thereby facilitating metabolic remodeling essential for cardiomyocyte maturation. In vivo, RF+Psr treatment significantly improved cardiac function and reduced fibrosis after MI compared to RF alone. CONCLUSIONPsr enhances direct cardiac reprogramming by activating PPAR signaling and promoting mitochondrial fission. These findings provide a novel mechanistic framework and suggest psoralen as a promising natural enhancer for cardiac regeneration strategies. Graphical abstract O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=193 SRC="FIGDIR/small/681288v1_ufig1.gif" ALT="Figure 1"> View larger version (40K): org.highwire.dtl.DTLVardef@13a5d27org.highwire.dtl.DTLVardef@b9cce5org.highwire.dtl.DTLVardef@3afef8org.highwire.dtl.DTLVardef@1398972_HPS_FORMAT_FIGEXP M_FIG C_FIG

molecular biology↗

FlexRibbon: Joint Sequence and Structure Pretraining for Protein Modeling

AO_SCPLOWBSTRACTC_SCPLOWProtein foundation models have advanced rapidly, with most approaches falling into two dominant paradigms. Sequence-based language models (e.g., ESM-2) capture sequence semantics at scale, and a number of recent works incorporate structural signals into sequence encoders. MSA-based predictors (e.g., AlphaFold 2/3) achieve accurate folding by exploiting evolutionary couplings, but their reliance on homologous sequences makes them less reliable in highly mutated or alignment-sparse regimes. We present FlexRibbon{ddagger}, a pretrained protein model that jointly learns from amino acid sequences and three-dimensional structures. Our pretraining strategy combines masked language modeling with diffusion-based denoising, enabling bidirectional sequence-structure learning without requiring MSAs. Trained on both experimentally resolved structures and AlphaFold 2 predictions, FlexRibbon captures global folds as well as flexible conformations critical for biological function. Evaluated across diverse tasks spanning interface design, intermolecular interaction prediction, and protein function prediction, FlexRibbon establishes new state-of-the-art performance on 12 different tasks, with particularly strong gains in mutation-rich settings where MSA-based methods often struggle.

molecular biology↗

The role of medicinal Cu2+ salts in expanding the multi-pathway application of tigecycline against drug-resistant bacteria

Tetracyclines are highly susceptible to photolysis, causing pollution in water, land and other ecosystems, as well as posing threats to human health. In this study, we found that only tigecycline and minocycline reduce the antibacterial activity in light, and this inferiority has severely hindered their scope of clinical use. By strategically regulating the photolysis pathway, we identified chalcanthite, which primarily contains Cu2+. By conducting antibiotic susceptibility test and electron microscopy analysis, Cu2+ salts (including copper sulfate, copper chloride, and copper gluconate) were confirmed to restore the antibacterial activity of tigecycline under light exposure against all tested bacterial strains. Spectroscopic characterization combined with quantum chemical calculations elucidated the molecular mechanism by which Cu2+ selectively modulates the photolytic degradation pathway of tigecycline and structural determinants (dimethylamino groups on the D ring). Multi-omics technologies revealed the mechanism underlying the Cu2+-mediated (non-antibacterial) inhibition of bacterial ferric citrate transporter. Animal infection models were developed using diverse drug-resistant pathogens to validate the in vivo therapeutic efficacy of copper gluconate-tigecycline combination therapy under photic conditions. These findings revealed effective adjuvants that facilitate the development of topical tigecycline formulations, and the establishment of a scientific foundation for synthesizing photostable tetracycline-based antibiotics through structural optimization.

biochemistry↗

Archetype analysis of lung adenocarcinoma premalignancy links heterogeneity in premalignant lesions to diverging features of invasive disease

ABSTRACTLung adenocarcinoma (LUAD), the most common form of lung cancer, is a heterogeneous disease with highly variable histopathologic features and clinical outcomes. LUAD premalignant lesions (PMLs) are localized proliferative lesions of atypical pneumocytes that proliferate along the alveolar walls. In this study, we characterized the molecular heterogeneity across PMLs, regardless of histologic designation, and identified four unique groups or "archetypes" of PMLs using bulk RNA and exome sequencing data from laser captured microdissected PMLs, normal, and tumor tissue from LUAD resection cases. The archetypes were defined based on expression of recurrent gene co-expression modules discovered using the LUAD PMLs we profiled and three publicly available datasets. One PML archetype, termed "proliferation", was associated with increased expression of genes involved in cell proliferation, a pro-tumor immune environment, and enrichment for EGFR driver mutations. Another archetype, termed "normal-like", was associated with similar features to normal tissue, an anti-tumor immune response, and lacked enrichment for driver mutations compared to other archetypes. We projected the PML archetypes into independent gene expression datasets profiling LUAD and found that tumors closest to the proliferation archetype were enriched for multiple features of aggressive LUAD, including high tumor grade and lymph node invasion, and had significantly lower disease-free survival. Proliferation archetype PMLs may represent a subset of lesions with more aggressive features that could improve patient stratification and suggest novel targets for lung cancer interception.

cancer biology↗

Degeneration and Impaired Resilience of Skull Bone and Hematopoietic Bone Marrow

Bone marrow health is central to transplantations, blood formation, and cancer progression. However, the bone marrow niche deteriorates with age, impairing haematopoietic stem cell function. Contrary to a recent report1 suggesting skull marrow resists ageing, our multi-laboratory investigation reveals the opposite: the skull marrow is among the vulnerable sites of age-related decline. Ageing skull niches consistently show loss of mesenchymal and osteoprogenitors, suppression of angiogenic and lymphatic programs, adipocyte accumulation, vascular senescence, DNA replication stress, mitochondrial dysfunction, cellular senescence, and heightened inflammation. Proteomic profiling further highlights this vulnerability, demonstrating that vertebral niches--unlike the skull--are relatively spared from these ageing hallmarks. Together, these convergent datasets overturn the notion of skull-specific resilience and instead establish the skull marrow as a fragile, degenerating environment. These findings redefine marrow ageing and highlight the skull as a critical, clinically relevant target for sustaining blood and immune health and reducing vulnerability to haematological disease.

cell biology↗

Directed evolution of a stem-helix targeting antibody enables MERS-CoV cross-neutralization through enhanced binding affinity

Broadly neutralizing antibodies (bnAbs) targeting conserved regions of the betacoronavirus spike are important for pan-betacoronavirus protection and pandemic preparedness. Here, we report on the isolation of a human monoclonal antibody, CC65.1, from a SARS-CoV-2 convalescent donor that targets the conserved S2 stem helix region. CC65.1 neutralizes various sarbecoviruses, including SARS-CoV-2, and binds the MERS-CoV spike but lacks MERS-CoV-neutralizing activity due to insufficient binding affinity. We utilized directed evolution to enhance the binding affinity of CC65.1 for the MERS-CoV S2 stem helix, yielding engineered antibody variants with newly acquired MERS-CoV-neutralizing activity. High-resolution structural analysis reveals key paratope mutations that optimize binding and stabilize epitope engagement. Our findings demonstrate the potential of rational antibody engineering to expand bnAb breadth across divergent betacoronaviruses. This work supports the development of engineered bnAbs and S2-targeted vaccines for broad betacoronavirus countermeasures and highlights strategies to achieve cross-lineage immunity for future pandemic threats. Author SummaryThe persistent emergence of new SARS-CoV-2 variants of concern that evade neutralizing antibody responses and other zoonotic betacoronaviruses with pandemic potential have provided strong motivation to develop broadly neutralizing antibodies (bnAbs) that target more conserved regions of the spike protein in sarbecoviruses and other betacoronaviruses. Here, we employed a directed evolution strategy to engineer the sarbecovirus-neutralizing antibody CC65.1, which targets the conserved S2 stem helix, to enhance its binding affinity with the MERS-CoV stem helix region, thereby conferring MERS-CoV neutralization. High-resolution structural studies of engineered CC65.1 revealed that key mutations reshape the paratope to better accommodate and stabilize the MERS-CoV S2 stem helix, resulting in increased binding affinity and neutralization potency. This study emphasizes the critical role of affinity maturation in expanding neutralization breadth and provides valuable insights for design of bnAbs to prevent and treat pandemic threats by betacoronaviruses.

immunology↗

Evolutionarily Conserved and Divergent Mechanisms of Dual Ca2+ Sensors in Synaptic Vesicle Exocytosis

Neurotransmitter release at the C. elegans neuromuscular junction is governed by a dual Ca{superscript 2} sensor system composed of SNT-1 and SNT-3, which are functional analogs of synaptotagmin-1 and -7 (Syt1/Syt7) in mammalian central synapses. In this study, we investigated how SNT-1 and SNT-3 mediate fast and slow neurotransmitter release through their potential interactions with the SNARE complex and their polybasic motifs. AlphaFold 3 models of SNT-1-SNARE and SNT-3-SNARE complexes accurately recapitulated the canonical Syt1 C2B-SNARE primary interface (Zhou et al., 2015, Nature) and precisely identified conserved binding residues within the C2B domains, as well as in SNAP-25 and Syntaxin, highlighting the evolutionary conservation of this interaction. Electrophysiological analyses using targeted mutagenesis demonstrated that both SNT-1 and SNT-3 require C2B-SNARE interactions and polybasic motifs within their C2 domains to drive evoked fast and slow neurotransmitter release. Notably, SNT-1 and SNT-3 exhibited differential dependence on distinct regions of the C2B-SNARE interface and their respective polybasic motifs, suggesting that Ca{superscript 2}-triggered fast and slow release operate via distinct mechanistic strategies. Furthermore, we found that SNT-1 mediates spontaneous neurotransmitter release through multiple pathways, involving not only the primary C2B-SNARE interface but also additional putative SNARE-binding interactions. Together, our findings uncover both conserved and divergent mechanisms for synaptic exocytosis regulated by the dual Ca2+ sensors in C. elegans.

neuroscience↗

CDK11 activates CDK12 to trigger the elongation of RNA Polymerase II

CDK11 is an essential cyclin-dependent kinase in higher eukaryotes, yet its specific substrates and functional roles have remained elusive. Our findings reveal that CDK11 is functionally analogous to the yeast Bur1 kinase, phosphorylating the repeat region of hSpt5 and the linker domain of Rpb1 in RNA Polymerase II (Pol II). Inhibition of CDK11 results in a significant reduction of active Pol II not only at transcription start sites but also along gene bodies. Further investigation finds that CDK11 is crucial for the phosphorylation and activation of CDK12, which is essential for the elongation of Pol II. Moreover, we find that CDK10 can partially compensate for the function of CDK11. Combining with earlier functional elucidation of CDK9 from us, the consequent roles of CDK7/8, CDK9, CDK10/11, and CDK12/13 in transcription regulation in metazoans are established.

biochemistry↗

Programming Microbe-Computer Interaction for Customized Life-Nonlife and Cross-Life Communication

This extension of human-computer interaction (HCI) paradigm to the non-intelligent microbe-computer interaction (MCI) is attractive for building generalized life-nonlife communication. Main challenges lie in how to construct biohybrid interface within individual MCI unit and thereafter program cross-life communication based on these MCI units. Here, we propose the design and construction of opto-interfaced MCI unit composed of engineered microbes and computer terminals, for enabling quantifiable transduction between microbial signals and digital signals. These units were able to serve as biohybrid sensors, actuators and transceivers, supporting biomonitoring, bio-actuation and microbe-centric bidirectional operations. Based on these functional units, we established programmable cross-life communication architectures capable of unicast, multicast, and logic-gate-driven communication with temporal dynamic control. The architectures were validated in biomedical scenarios including tumor hypoxia biomarking, epidermal heavy metal detoxification, and colitis intervention, demonstrating robust performance in closed-loop diagnostic-therapeutic applications. Additionally, we integrated MCI with an unmanned surface vessel for environmental monitoring, highlighting its cross-domain potential. Collectively, we provide a scalable strategy for building customized life-nonlife and cross-life communication, which is expected as an important complement to internet of things (IoT).

synthetic biology↗

Observing biological spatio-angular structures and dynamics with statistical image reconstruction and polarized fluorescence microscopy

Understanding molecular orientation and density distributions is essential for exploring biological structure and function. Polarized fluorescence microscopy (PFM) provides insights into molecular architecture but struggles to resolve three-dimensional (3D) molecular orientation distributions, particularly in densely labeled or structurally complex specimens. To address this, we introduce the efficient generalized Richardson-Lucy (eGRL) algorithm, a robust framework for reconstructing 3D molecular density and orientation (spatio-angular) distributions from PFM data. By modeling the imaging process in spatio-angular hyperspace, we propose a maximum-likelihood solution enhanced by dimensionality reduction and angular domain transformation to overcome computational challenges. eGRL improves accuracy and efficiency across different PFM implementations, enabling use on standard platforms. We utilize our methods to resolve biological spatio-angular structures and dynamics otherwise impossible to resolve, including the tangential alignment of actin filaments in U2OS cells, nanowire-guided cytoskeletal organization in NIH3T3 cells, rotational actin patterns in live HeLa protrusions, and membrane tension-induced anisotropy in live macrophages.

bioengineering↗

Illuminating the Virosphere's Dark Matter using Hierarchical Deep Learning

Systematic discovery of novel viruses is essential for pandemic preparedness, understanding tumor-associated viruses, developing viral delivery systems, and advancing biomedical applications. Yet, the majority of sequences in metagenomic datasets lack close relatives in existing references, representing a vast viral "dark matter" whose biology and evolution remain largely unknown. The central task is threefold: 1) to determine whether a genome is viral or non-viral, 2) to correctly assign viral genomes to known lineages when possible, and, critically, 3) to recognize when no existing lineage applies and thereby identify candidates for entirely novel viral groups. Existing approaches, which depend on sequence homology or narrow markers, struggle to capture this uncharted viral space. Here we present DeepVirus, a hierarchical transformer-based framework that models viral genomes as structured sequences of protein-coding genes. By combining protein-level embeddings from a foundation model with genome-aware representations, DeepVirus not only achieves accurate classification across deep taxonomic hierarchies, but also extends beyond conventional classification to detect and organize candidate novel viral lineages through open-set recognition. Applied to large-scale metagenomic resources, DeepVirus uncovered extensive viral diversity, including previously uncharacterized RNA-dependent RNA polymerases (RdRps), thereby expanding the known evolutionary space of RNA viruses. DeepVirus integrates deep learning with genome-aware open-set discovery to illuminate viral dark matter, providing a foundation for systematic viral taxonomy and advancing exploration of the global virosphere, with broad implications for safeguarding human health.

bioinformatics↗

Molecular architecture of OXGR1 reveals evolutionary conserved mechanisms for metabolite surveillance

The ability of cells to sense and respond to metabolic signals is fundamental to life, yet the molecular mechanisms underlying metabolite surveillance remain incompletely understood. Here, we elucidate the structural basis of metabolite recognition by OXGR1, a GPCR that monitors key intermediates in the tricarboxylic acid cycle (TCA). Using cryo-electron microscopy, we determined four cryo-EM structures of OXGR1 bound to -ketoglutarate (AKG), itaconate (ITA), and structural related succinate (SUC) and maleate (MA). These structures reveal a positively charged binding pocket and an extensive hydrogen-bond network critical for OXGR1 recognizing dicarboxylic acids. In addition, we identify a distinct arrangement of hydrophobic residues that modulates ligand potency and selectivity. Mutational analysis and molecular dynamics simulations further demonstrate that non-canonical micro-switch motifs, including FRY and NLxxY, are essential for ligand recognition and receptor activation. Comparative structural and evolutionary analyses indicate that these mechanisms are conserved across species, underscoring the critical role of OXGR1 in maintaining metabolic homeostasis. Together, our findings provide a mechanistic framework for metabolite sensing via OXGR1 and suggest potential strategies for therapeutic modulation of metabolic and inflammatory diseases.

molecular biology↗

DNA Data Storage Architecture via Ligation of Dynamic DNA Bytes

The explosive growth of digital data is overwhelming conventional storage media, creating an urgent need for more efficient solutions. DNA offers immense potential for digital data storage, yet most systems remain static and archival. Here, we present a modular DNA storage architecture based on dynamic DNA bytes (DynaBytes)-- pre-fabricated DNA segments that can be ligated into reconfigurable information units. Utilizing core, functional and control DynaBytes, we stored 210,776 bits (26,347 bytes) of digital information organized within a file-system, and demonstrated CRUD (Create-Read-Update-Delete)-like operations, hierarchical access and nanopore-based realtime retrieval. Robust data recovery was achieved under [~]100x error-prone sequencing using streamlined error correction and fuzzy decoding. By relying on in vitro ligation of standardized components, DynaBytes reduces cost, scales efficiently, and allows interactive, rewritable storage. These features advance DNA storage beyond passive archiving toward a reconfigurable framework, opening new possibilities for dynamic, practical and large-scale DNA-based data systems.

synthetic biology↗

Large-language-model-based antibiotic resistance gene prediction and resistomes mining in cyanobacterial blooms

The rapid spread of antibiotic resistance genes (ARGs) in aquatic ecosystems poses a serious public health threat. Conventional ARG detection methods, including sequence-alignment and machine learning, are limited by high false-negative or false-positive rates, especially due to sequence diversity and class imbalance in metagenomic data. Moreover, few tools offer comprehensive solutions for ARG identification, classification, and functional annotation. To address these limitations, we developed ESMARG, a novel protein language model framework based on ESM1v, for accurate ARG identification, classification, and annotation. Trained with both ARG and abundant non-ARG sequences from real-world metagenomes to enhance robustness and reduce false positives, ESMARG significantly outperformed traditional alignment-based (BLAST, DIAMOND) and recent deep learning models (ARGNet, ARG-SHINE), achieving 0.998 precision, 0.939 recall, and 0.968 F1-score for ARG identification. Functional and mechanism classification modules also demonstrated high accuracy (0.986 and 0.987) and computational efficiency. We applied ESMARG to analyze 26 cyanobacterial aggregate (CA) metagenomes collected from Lake Taihu across a full annual cyanobacterial harmful algal bloom (CyanoHAB) cycle. A total of 110 ARGs, spanning 24 drug classes and multiple resistance mechanisms, were detected, with seasonal shifts in ARG abundance and composition. Strong correlations were identified between CA resistomes and microbial community structure, especially with cyanobacteria. Further, ARG abundance was shaped by both biotic factors--such as cyanobacterial dominance and bacterial functional profiles--and abiotic variables, including biochemical oxygen demand and water temperature. Our findings demonstrate the power of ESMARG for high-resolution ARG profiling and reveal complex ecological interactions between resistomes, microbiomes, and environmental factors in CyanoHAB. HighlightsO_LIDeveloped ESMARG model outperforms BLAST/other deep learning models in ARG detection C_LIO_LIARG showed seasonal dynamics in cyanobacterial metagenomes across algal bloom cycle C_LIO_LIARG was strongly correlate with microbiome structure and cyanobacteria dominance C_LIO_LIARG abundance was shaped by cyanobacteria, bacterial functions, BOD and temperature C_LI

bioinformatics↗